GPT-6 Astra scored 14% on MazeBench, a user reports
The user says Astra spent 60-plus hours on the test, described as a spatial reasoning evaluation in a 3D open world.
TLDR
GPT-6 Astra spent 60-plus hours in MazeBench and finished with a 14% score, according to a user who describes the test as a 3D open-world spatial reasoning evaluation.
Combined views
15
1 Source, first seen 23d ago
2 reposts